What Would Your Rating Be if You Stopped Making Stupid Mistakes?
Have you ever played a chess game you were really pleased with?You know, when everything comes together for a change, and you think you've cracked it.
Then you play again and it's dire. A game ruined by a stupid move.
Well, this happened to me the other day (again). So, I decided to investigate at what level players stop making stupid moves.
(By 'stupid moves' I mean moves clearly below your level. Taking a piece that's defended. Hanging your Queen to a one-mover. Allowing mate in one when it was easily prevented.)
So, I took to the Lichess database.
Method: I decided to look at 5+3 blitz games, with the caveat that I would ignore time scrambles. I would only look at moves when the players had plenty of time on their clocks and made an obvious silly mistake, as listed above.
2100?
My first sample of players was 2100+ under the conditions above. The very first game I looked at featured a player rated 2150. I flicked through the game looking for major blunders and found that he'd taken a pawn with a bishop that was defended by a knight. Absolutely no compensation. A piece for a pawn. Game lost. I sampled a few more at this level and the result was the same. This was not expected, as I'm regularly playing 2000-2100+ on Lichess 5+3.
2300?
So, I went up to 2300. Same controls and conditions. I found a player at 2300 completely at random who plays 5+3. First game. A Queen move with 4 minutes on the clock walks into a skewer. Drops a rook. Game over. I tried more games at 2300 and found the same thing. Losses due to basic errors. Time to move up.
2400?
Here I thought, surely lost games come from inaccuracies or blunders in time pressure. The first game I looked at was an early middlegame, and the 2400+ player allowed a simple knight fork, winning a rook for a knight. Game over. I tried more games at 2400, and some of them did not contain such errors, but I would say one-third of them did.
2500?
At this level, I found that only 1 in 11 or 1 in 12 games contained such silly errors. So, the answer to my original question is 2500. At least in blitz. Maybe now I'll look at rapid and classical.
However, in my brief investigations here I also found something extremely interesting. All these silly mistakes I found had one thing in common – the moves were all played very quickly. Of course, this is blitz, but I'm talking about an average of two seconds each. That's two seconds when the players often had over four minutes on their clocks remaining. In other words, they played too quickly, even within the constraints of the time control.
There's no way a 2400 doesn't know what a knight fork is, so reducing these silly errors has to be about taking more care in every position. This got me thinking about how to approach games in the future: taking more time on every move. Even just a few more seconds holding back to double-check that each move is safe could dramatically reduce these types of errors in the future.
(The short video accompanying this blog post can be found in the forum pages.)